Bayes' Rule Goes Quantum: A 250-Year-Old Theory Learns New Tricks

Abstraction: Quantum Bayes' rule derived from minimum-change principle matches Petz recovery map

Key points:

• Researchers at NUS (Valerio Scarani), HKUST (Ge Bai), and Nagoya University (Francesco Buscemi) derived a quantum analogue to Bayes' rule from the principle of minimum change; published in Physical Review Letters (DOI: 10.1103/5n4p-bxhm)
• Classical Bayes' rule updates beliefs with minimum change to the joint probability distribution; the quantum version minimizes change in terms of quantum fidelity between forward and reverse processes
• Their derivation yields the Petz recovery map (Dénes Petz, 1980s) — first time it has been derived from a fundamental principle rather than identified by its properties alone
• Quantum fidelity measures closeness between quantum states; maximizing fidelity equals minimizing state change after measurement
• Petz map has potential applications in quantum error correction and quantum machine learning
• Classical Bayes' rule is used in medicine, meteorology, and ML; quantum version extends this reasoning to quantum systems where measurement collapses state probability

Connections: Bayesian Inference · Quantum Computing · Quantum Error Correction · Quantum Machine Learning

Source: https://scitechdaily.com/bayes-rule-goes-quantum-a-250-year-old-theory-learns-new-tricks/